INDEX

Explanations

instances of phrases indicating conditions that diminish effectiveness or value

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alta

-0.09

ãģ®æĸ¹

-0.08

Î

-0.07

/gtest

-0.07

UDGE

-0.07

leness

-0.07

hev

-0.07

UnderTest

-0.06

reck

-0.06

-required

-0.06

POSITIVE LOGITS

 efforts

0.11

try

0.11

 tried

0.11

 tries

0.10

 attempted

0.10

 attempts

0.10

 effort

0.10

 attempt

0.09

Try

0.09

 Attempts

0.09

Activations Density 0.032%